[nojs]

This looks very useful.

I would love something similar that lets me plug in actual Claude Code/Codex with their original agent loop, prompting etc, and just handles the multiplexing, worktrees, isolation, etc automatically (it looks like this tool doesn’t support that). Because I think a lot of the power of eg CC comes from the engineering they’ve done to the tool rather than the underlying model.

How are people doing this at the moment?

[Coolin96]

I’ve been using (and loving) https://github.com/raine/workmux which brings together tmux, git worktrees, and CLI agents into an opinionated workflow.

[jadbox]

Oh this looks perfect. This feels like the Linux way to keep tools separated to their primary function. I don't want my MUX tool to do AI stuff, as I have other (better) tools meant for that.

[nojs]

Thanks for posting this - just set it up and liking it so far

[widenrun]

I'm using conductor.build and running both Claude Code and Codex. It's taken 95% of my workflow and I'm loving it (I'm not affiliated with them at all, genuinely enjoying it and hoping it succeed)

[scottmf]

Same here. Codex support is a recent addition however, and it’s not clear if MCP servers and other rules apply to Codex. Also it would be nice to be able to just have a session working on the main branch as concurrent work in worktrees can get messy

[CuriouslyC]

Amazing to me all these apps duplicating well tuned Github functionality.

[ttobi]

I have build my own tool for this https://github.com/tobias-walle/agency It uses tmux to run the agents and offers some convenience commands that, e.g. lets you merge the changes back into its original branch. I added some simple idle and change detection, so you can see which agent needs your attention. As all agents are just simple cli commands it is very easy to extend the config with your tool of choice.

[ammario]

Author of Mux here:

I started building it out that way but found it very challenging to create parity between the models. E.g. they have different tools, system prompts, interruption semantics, cost tracking etc. The spirit of the product is decoupling the LLM from the UI, so we went with a custom loop / tools that can perform decently across all models.

[CuriouslyC]

Since CC is a terminal app you can spin up a container with your project, dependencies and CC already installed easily via docker, then just ssh in and run claude remotely. Tmux if desired. Save work by doing pull requests.